Thanks to biotechnology, immunotherapy has become standard of care along many a cancer patient’s journey, with many targeted drugs now available. One…
Multi-cancer Early Detection Blood Tests (MCED) Debut
A 52-year-old woman is at her annual physical exam. The physician assistant mentions he’ll need two extra vials of blood for new cancer screening tests, one just FDA-approved, the other available as part of a clinical trial.
“But I already get mammograms and colonoscopies based on family history, and my husband gets his PSA screen for prostate cancer every year. So far, so good. Why do I need these new tests?” the patient asks.
“They can catch cancers much earlier, from DNA and proteins in your blood plasma, the liquid part. Including cancers much rarer than breast, colon, and prostate.”
“Sure,” says the patient, rolling up a sleeve. She’d be one of the first to have “multi-cancer early detection” – MCED – blood tests that zero in on clues that cancer cells shed into the bloodstream. A treatment begun early is more likely to work. An MCED blood test could be a gamechanger for people who haven’t had cancer.
Liquid Biopsy
Cancer accounts for 1 in every 6 deaths; only about 40 percent are detected early enough to treat. A mammogram found my breast cancer early; a doctor who noticed a bump in my throat found thyroid cancer.
For those who’ve had cancer, a powerful tool to detect return is a “liquid biopsy,” in which a body fluid – typically blood, but also possibly urine, sputum, or feces – is probed for tiny pieces of DNA that harbor mutations that cause cancer. Research on liquid biopsies began three decades ago.
The snippets are called “cell-free circulating DNA,” or cfDNA. Overlapping the pieces can reassemble the entire genome sequence of cancer calls plucked from the bloodstream, enabling identification of mutations. In addition, the expression of cancer genes (whether genes are turned on or off) can be deduced from the patterns of methyl groups (a carbon bound to three hydrogens) that cling to the DNA. In cancer, a tumor suppressor gene may be cloaked in methyls and silenced, lifting that suppression. Or an oncogene may be activated, causing cancer, when methyls fall off.
A liquid biopsy – just a blood test – is much less painful and invasive than a traditional surgical biopsy, which samples cancer cells from a solid tumor. And tumor DNA is more specific than a protein biomarker, which may also be present on healthy cells but is more abundant on cancer cells.
Liquid biopsies so far have focused on people who’ve already had cancer, spotting recurrence or monitoring response to treatment.The Holy Grail is to use it in people who haven’t had cancer, nor likely to undergo other screens, such as PSA, due to a risk factor like age or family history. Ovarian cancer is the classic example. It’s often diagnosed late because the telltale bloating and fatigue are vague and common, easily attributed to something innocuous, like a change in diet or exercise routine.
MCED blood tests are most promising for the many rare cancers that often go undetected far too long. They also will likely improve on older screening methods for sensitivity – identifying individuals who have cancer – and specificity – excluding those who do not have cancer.
The new MCED blood tests, one type approved, another experimental, detect DNA or proteins.
Telltale DNA
Already available by prescription is “multi-cancer early detection from a single blood draw,” from healthcare company GRAIL. In October 2023, the company published findings of their Pathfinder study in The Lancet. It followed 6,662 adults over age 50, without symptoms, who had taken the test.
Checking methylation patterns (whether cancer genes are turned on or off) more than doubled the number of new cancers detected, with nearly half in early stages. After follow-up based on the specific cancers, most diagnoses came in under three months, far faster than using standard screening tests.
The DNA MCED test, called Galleri, covers more than 50 types of cancer, including many not currently found with recommended screening tests. It spots three types of bile duct cancers, and cancers of the small intestine, oral cavity, vagina, appendix, penis, peritoneum, and others, as well as the common ones such as pancreas, prostate, and bladder cancers.
Machine learning is used to link DNA methylation patterns to useful clinical information, explained in a report in Cancer Cell. The approach reveals cancer type, organ of origin, and distinguishes cancer cells from normal ones that are just old and shedding DNA.
The test identified 36 cancers in 35 participants, with one person diagnosed with two cancers. Standard cancer screenings identified 29 cancers.
Probing Proteins
Like liquid biopsy for DNA, the power of identifying the set of proteins in blood plasma comes from surveying many bits of information.
Screening tests that detect certain proteins associated with increased risk of specific cancers are familiar: PSA for prostate cancer, estrogen and progesterone receptors on breast cancer cells to choose treatment, and elevated levels of CA125 to indicate ovarian cancer.
But these are a tiny sampling of what’s possible by cataloging levels of many proteins in blood plasma – a proteomics approach.
A “proof-of-concept” study just published in BMJ Oncology made headlines: “Sex-specific panel of 10 proteins can detect 18 different early-stage cancers.”
Researchers at Novelna Inc. in Palo Alto deduced protein-based “signatures of solid tumors” in specific organs from plasma samples from 440 people diagnosed with 18 different types of cancer before treatment, and from 44 healthy blood donors. They measured the abundance of more than 3,000 proteins, noting the highs and lows associated with cancer types. The proteins chosen to survey were based on the biochemical pathways they are part of and that are subverted in cancer, such as signal transduction, cellular adhesion, and cell division.
A key part of the study was including biological sex, based on the fact that certain cancers are much more prevalent in males or females. Again, AI helped to identify site of origin as well as distinctions within a body part, such as medullary versus follicular versus papillary thyroid cancer, and small cell and non-small cell lung cancers.
Low levels of ten plasma proteins emerged as meaningful predictors of 18 types of solid tumors, tracking with sex. Levels of 150 proteins indicated site of origin, and the protein signature also indicated stage, even very early.
In an editorial accompanying the BMJ Oncology report, Holli Loomans-Kropp, of Ohio State University pointed out that “Demonstrable sex-specific differences in cancer—including age of onset, cancer types and genetic alterations—suggest this approach would be useful. Incidence of larynx, throat and bladder cancers are higher in men, anal and thyroid in women. Cancers due to mutations in the p53 gene begin earlier in women. And different types of mutations drive acute lymphoblastic leukemia in men and women.”
Other genetic measures differ between the sexes, such as number of mutations, copy number alterations, and methylation patterns. X chromosome genes suppress tumor growth, and we females of course have two Xs compared to the male of the species’ one.
In the future, perhaps a multi-cancer early detection blood test will become as routine as a cholesterol check. Finding cancer early, so easily, promises to ultimately save many lives!
I am just starting this journey as part of Observe versus Chemotherapy for Uterine Serous Carcinoma. Supposedly surgery has removed this aggressive cancer completely, but just in case, I feel blessed to have these DNA tests available for an earlier notice of recurrence.
I suspect that this sort of test may turn out to be problematic in some settings, particularly in the detection of cancers that are extremely common and often clinically insignificant, particularly prostate cancer in older men and papillary carcinoma of the thyroid. I would approach such detections with great caution.
That’s an interesting point – how often those early detections pan out will need to be determined, and the characteristics of those in whom they are informative better understood. Thanks for posting!
I just read an article on this topic in Science. Approximately 62% of detections using this method appear to have been false positives. Only 0.2% of the initially screened population had a previously undetected, early stage, potentially curable cancer.
There is the potential to do a great deal of harm by deploying this technology widely. At this point, I don’t think that this technology should be widely used for screening,